“ Instead of teaching people to just blindly accept what the AI spits out, we force them to practice what we call harness engineering,” Shayan goes on.“ We build a culture of relentless systems design and testing. You develop critical thinking and judgment by making people responsible for the intent of the system.”
The rise of cognitive debt Coding itself has changed markedly with the arrival of AI assistants, tools that generate or suggest software code automatically.
Shayan highlights a risk he and Thoughtworks term“ cognitive debt”:“ AI tools are phenomenal accelerators and developers genuinely love using them. But teams are shipping code significantly faster than they actually understand it.”
The result is a“ dangerous gap” between what AI produces today and what humans can maintain in the future. Software engineering is, therefore, shifting away from manually typing logic towards supervising and constraining AI systems effectively. Today’ s engineers are increasingly acting as reviewers, orchestrators and architects.
aimagazine. com 25